3:04 a.m.
a.lobianco(a)univpm.it wrote 2008-04-14 09:38:29 :
Hello all.. I'm wandering if it is possible or scheduled for
future
releases to read Excel data files from the gretl command line with
option to specific the sheet name and with the option to read datasets
that are placed across rows instead of columns..
You can place a feature request here:
http://sourceforge.net/tracker/?atid=416806&group_id=36234&func=b...
Andreas
8:35 p.m.
New subject: Request: options in reading excel data files
On Wed, 16 Apr 2008, andreas.rosenblad(a)ltv.se wrote:
a.lobianco(a)univpm.it wrote 2008-04-14 09:38:29 :
> Hello all.. I'm wandering if it is possible or scheduled for future
> releases to read Excel data files from the gretl command line with
> option to specific the sheet name and with the option to read datasets
> that are placed across rows instead of columns..
You can place a feature request here:
http://sourceforge.net/tracker/?atid=416806&group_id=36234&func=b...
Yes, that's right. However, in response to the original question,
we don't currently have any plans to beef up the XLS importer. We
will, of course, hasten to fix any bugs that are reported.
As a general comment, please note that the CSV/text importer has
some refinements that are not present in the XLS importer. So the
preferred import route, if the data file contains anything
"non-standard", is via a CSV/text export from Excel, rather than
directly XLS -> gretl.
Also note that the "dataset transpose" command is available within
gretl for imported data that are the wrong way round (variables in
rows rather than columns).
Out of ignorance, and as a matter of interest: has anyone compared
the functionality of the importation of XLS data into gretl, with
the importation of such data into other econometrics programs?
I.e., how do we compare with Eviews, ox, R, or whatever, in terms
of flexibility (offsets, choice of worksheet), detection of
variable names and time-series dating, and so on?
Allin Cottrell
1:44 p.m.
New subject: Request: options in reading excel data files
It has been nearly 5 years since I have used Eviews. But the import from XLS
was so bad back then that I wrote a several page tutorial for other members
of my econometrics class. I understand that the tutorial is still
being circulated.
As a point of reference, I have had no complaints from student regarding
import of data into GRETL from any of the 5 econometric classes that I
have taught... but we only import data that has been spit out of Excel into
tab, space, or command delimited format.
Peter
At 06:35 PM 4/16/2008, you wrote:
On Wed, 16 Apr 2008, andreas.rosenblad(a)ltv.se wrote:
> a.lobianco(a)univpm.it wrote 2008-04-14 09:38:29 :
>
> > Hello all.. I'm wandering if it is possible or scheduled for future
> > releases to read Excel data files from the gretl command line with
> > option to specific the sheet name and with the option to read datasets
> > that are placed across rows instead of columns..
>
> You can place a feature request here:
> http://sourceforge.net/tracker/?atid=416806&group_id=36234&func=b...
Yes, that's right. However, in response to the original question,
we don't currently have any plans to beef up the XLS importer. We
will, of course, hasten to fix any bugs that are reported.
As a general comment, please note that the CSV/text importer has
some refinements that are not present in the XLS importer. So the
preferred import route, if the data file contains anything
"non-standard", is via a CSV/text export from Excel, rather than
directly XLS -> gretl.
Also note that the "dataset transpose" command is available within
gretl for imported data that are the wrong way round (variables in
rows rather than columns).
Out of ignorance, and as a matter of interest: has anyone compared
the functionality of the importation of XLS data into gretl, with
the importation of such data into other econometrics programs?
I.e., how do we compare with Eviews, ox, R, or whatever, in terms
of flexibility (offsets, choice of worksheet), detection of
variable names and time-series dating, and so on?
Allin Cottrell
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
--
No virus found in this incoming message.
Checked by AVG.
Version: 7.5.519 / Virus Database: 269.22.12 - Release Date:
4/10/2008 12:00 AM
3:24 p.m.
New subject: Request: options in reading excel data files
On Thu, 17 Apr 2008, Peter Davidoff wrote:
It has been nearly 5 years since I have used Eviews. But the
import from XLS was so bad back then that I wrote a several page
tutorial for other members of my econometrics class. I
understand that the tutorial is still being circulated.
As a point of reference, I have had no complaints from student
regarding import of data into GRETL from any of the 5
econometric classes that I have taught... but we only import
data that has been spit out of Excel into tab, space, or command
delimited format.
Thanks for the info. I haven't used current Eviews, but the last
time I tried importing into Eviews from XLS it struck me as a
terrible pain, consistent with what you're saying.
Allin Cottrell
10:50 a.m.
New subject: Request: options in reading excel data files
EVIEWS has more options, and their import from an XLS file is presented to
the user as basically the same as if a CSV or simple text file was being
imported. Other formats can be converted using very similar steps or
commands.
The GRETL Excel import gets the job done, but I suspect many would prefer
the EVIEWS way of showing a single screen (as opposed using 'Forward') to
set all options and not insisting that row 1 has variable names.
See
http://wps.aw.com/wps/media/objects/3254/3332253/tutorials/eviews_tutorial_4
_1.pdf for a 'import/read' screen shot for EVIEWS 4.1
and for EVIEWS 6.0 see http://www.eviews.com/illustrated/Illustrated.html
and Chapter 2 for 'EVIEWS-Meets-Data.pdf' for newer bells and whistles that
include 'guessing' each column's format as text, number or date.
Finally, I wanted to be able to automate importing a set of similar EXCEL
files in GRETL and could not see how to get full control with something like
open excel 'C:\data.xls' 1 1 'sheet1'
I think the GUI steps should produce the full set of options, as opposed to
the command log showing just
open "C:\test1.xls"
Still GRETL is great and clearly a top choice for students. I especially
like the command log feature that documents each GUI click-on step. And the
interface to R is a great idea that opens it up to more researchers.
Allin-- As I work with it more I will give a list of feature requests on
the data side, such as an ODBC or MySQL interface, but these may go farther
from your original vision. For general import/export I suspect that making
use of XML as the intermediate format between any two packages would be the
best long run development course, especially since EXCEL 2007 writes XML as
the default.
-----Original Message-----
From: gretl-users-bounces(a)lists.wfu.edu
[mailto:gretl-users-bounces@lists.wfu.edu] On Behalf Of Allin Cottrell
Sent: Thursday, April 17, 2008 3:24 PM
To: Gretl list
Subject: Re: [Gretl-users] Request: options in reading excel data files
On Thu, 17 Apr 2008, Peter Davidoff wrote:
It has been nearly 5 years since I have used Eviews. But the
import from XLS was so bad back then that I wrote a several page
tutorial for other members of my econometrics class. I
understand that the tutorial is still being circulated.
As a point of reference, I have had no complaints from student
regarding import of data into GRETL from any of the 5
econometric classes that I have taught... but we only import
data that has been spit out of Excel into tab, space, or command
delimited format.
Thanks for the info. I haven't used current Eviews, but the last
time I tried importing into Eviews from XLS it struck me as a
terrible pain, consistent with what you're saying.
Allin Cottrell
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
11:45 p.m.
New subject: Request: options in reading excel data files
On Fri, 18 Apr 2008, Michael Boldin wrote:
Finally, I wanted to be able to automate importing a set of
similar EXCEL files in GRETL and could not see how to get full
control with something like
open excel 'C:\data.xls' 1 1 'sheet1'
In CVS and the Windows snapshot you now have, for example:
open foo.xls --sheet=1 --xoffset=0 --yoffset=0
This applies to XLS, gnumeric and ODS. Zero or more of the three
new parameters may be given. The defaults are shown above: sheet
number 1, x (column) and y (row) offsets both zero. The sheet is
specified as a positive integer and the offsets as non-negative
integers.
Allin Cottrell
6:06 p.m.
New subject: Request: options in reading excel data files
Allin
I tried from
http://sourceforge.net/project/showfiles.php?group_id=36234&package_i...
the Windows EXE
gretl-1.7.4.exe Mirror 9954757 5432 i386 .exe (32-bit Windows)
But it did to seem to use the -- arguments with open
Here is what my script output shows, for the attached 'opentest1.inp' and
XLS files. All --xoffset= and --yoffset= values give the same results as =0
gretl version 1.7.4
Current session: 2008/04/21 18:03
? open 'C:\Documents and Settings\Michael\Desktop\test1.xls' --sheet=1 \
--xoffset=2 --yoffset=1
Listing 3 variables:
0) const 1) X1 2) X2
? summary X1
Summary Statistics, using the observations 1 - 5
for the variable 'X1' (5 valid observations)
Mean 3.0000
Median 3.0000
Minimum 1.0000
Maximum 5.0000
Standard deviation 1.5811
C.V. 0.52705
Skewness 0.00000
Ex. kurtosis -1.3000
? summary X2
Summary Statistics, using the observations 1 - 5
for the variable 'X2' (5 valid observations)
Mean 300.00
Median 300.00
Minimum 100.00
Maximum 500.00
Standard deviation 158.11
C.V. 0.52705
Skewness 0.00000
Ex. kurtosis -1.3000
Gretl 'About' tells meI am using 1.7.4 and build of 3-20-2008
Do I need to compile from the C code?
--Michael Boldin
-----Original Message-----
From: gretl-users-bounces(a)lists.wfu.edu
[mailto:gretl-users-bounces@lists.wfu.edu] On Behalf Of Allin Cottrell
Sent: Friday, April 18, 2008 11:46 PM
To: Gretl list
Subject: RE: [Gretl-users] Request: options in reading excel data files
On Fri, 18 Apr 2008, Michael Boldin wrote:
Finally, I wanted to be able to automate importing a set of
similar EXCEL files in GRETL and could not see how to get full
control with something like
open excel 'C:\data.xls' 1 1 'sheet1'
In CVS and the Windows snapshot you now have, for example:
open foo.xls --sheet=1 --xoffset=0 --yoffset=0
This applies to XLS, gnumeric and ODS. Zero or more of the three
new parameters may be given. The defaults are shown above: sheet
number 1, x (column) and y (row) offsets both zero. The sheet is
specified as a positive integer and the offsets as non-negative
integers.
Allin Cottrell
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
6:12 p.m.
New subject: Request: options in reading excel data files
On Mon, 21 Apr 2008, Michael Boldin wrote:
I tried from
http://sourceforge.net/project/showfiles.php?group_id=36234&package_i...
the Windows EXE
gretl-1.7.4.exe Mirror 9954757 5432 i386 .exe (32-bit Windows)
But it did to seem to use the -- arguments with open...
Thanks for testing! But for this new feature you need the current
snapshot:
http://ricardo.ecn.wfu.edu/pub/gretl/gretl_install.exe
It'll be in release 1.7.5 when that happens, which ought to be
fairly soon.
Allin.
8:17 p.m.
New subject: Request: options in reading excel data files
I got the right download now with a April build date (sorry I somehow
misread a 3-20-2008 build as last week)
In my first tests, the XLS imports worked with the different --xoffset and
--yoffset options. But I then tried a script with --sheet=2 and this did
not work
My first guess was the --sheet=1 option opens the last open XLS sheet (as
saved in EXCEL), even if is not named 'Sheet1', and the options you added
count from the default sheet and =2 or =3 might not exist.
But that does not explain all of the errors I get, as now I can not import
any XLS files, even through the menu.
The error is always
First character of varname (0x0) is bad (first must be alphabetical)
Even other spreadsheets that the old 1.7.4 can read can not be read by the
new Gretl now. But I can open any Gretl sample dataset.
Can I have two Gretl versions installed?
-----Original Message-----
From: gretl-users-bounces(a)lists.wfu.edu
[mailto:gretl-users-bounces@lists.wfu.edu] On Behalf Of Allin Cottrell
Sent: Monday, April 21, 2008 6:13 PM
To: Gretl list
Subject: RE: [Gretl-users] Request: options in reading excel data files
On Mon, 21 Apr 2008, Michael Boldin wrote:
I tried from
http://sourceforge.net/project/showfiles.php?group_id=36234&package_i...
the Windows EXE
gretl-1.7.4.exe Mirror 9954757 5432 i386 .exe (32-bit Windows)
But it did to seem to use the -- arguments with open...
Thanks for testing! But for this new feature you need the current
snapshot:
http://ricardo.ecn.wfu.edu/pub/gretl/gretl_install.exe
It'll be in release 1.7.5 when that happens, which ought to be
fairly soon.
Allin.
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
11:51 p.m.
New subject: Request: options in reading excel data files
On Mon, 21 Apr 2008, Michael Boldin wrote:
In my first tests, the XLS imports worked with the different
--xoffset and --yoffset options. But I then tried a script with
--sheet=2 and this did not work
What happened?
My first guess was the --sheet=1 option opens the last open XLS
sheet (as saved in EXCEL), even if is not named 'Sheet1', and
the options you added count from the default sheet and =2 or =3
might not exist.
For gretl's purposes the sheets are simply numbered from 1 to N,
in the order in which they appear in the OLE file. I don't think
that order would be changed depending on which sheet had been
active when the file was last saved. (I know that's not the case
when you edit an XLS file using gnumeric.)
But that does not explain all of the errors I get, as now I can
not import any XLS files, even through the menu.
What has changed since you successfully imported using the
--xoffset and --yoffset options?
At this point I've done a fair amount of testing on on Linux but
not on Windows; I'll need to try that. But the code is common, so
I'm puzzled.
Can I have two Gretl versions installed?
No, that won't work. There will only be one set of registry
entries for gretl, so whichever was installed last will (I
suppose) be the only one to work.
Allin Cottrell
7:47 a.m.
New subject: Request: options in reading excel data files
> In my first tests, the XLS imports worked with the different
> --xoffset and --yoffset options. But I then tried a script with
> --sheet=2 and this did not work
What happened?
With the second (new) GRETL, I get either a MS Windows error, saying
'Gretl32.win.exex encountered a problem and needs to shut down'
whenever I import an .xls file or a Gretl box error 'First character of
varname (0x0) is bad (first must be alphabetical)'
What has changed since you successfully imported using the
--xoffset and --yoffset options?
Nothing I was aware of. I might have accidentally loaded the first Gretl
however, along with the new one. I tried re-booting with no luck.
> Can I have two Gretl versions installed?
No, that won't work. There will only be one set of registry
entries for gretl, so whichever was installed last will (I
suppose) be the only one to work.
Both seem to work in other respects (other than XLS importing) but I will
re-install only the new one and tell you what happens.
12:08 a.m.
New subject: Request: options in reading excel data files
On Fri, 18 Apr 2008, Michael Boldin wrote:
The GRETL Excel import gets the job done, but I suspect many
would prefer the EVIEWS way of showing a single screen (as
opposed using 'Forward') to set all options and not insisting
that row 1 has variable names.
Just one more observation for now: we don't require that row 1 has
variable names: we fill in "v1", "v2", etc., if we don't find any
names. This has been the case for quite a long time.
Allin.
5940
days inactive
5946
days old
11 comments
4 participants
participants (4)
-
Allin Cottrell -
andreas.rosenblad＠ltv.se -
Michael Boldin -
Peter Davidoff